Data format requirements

GraphHelix accepts CSV, Excel, SPSS, and Stata files. This page covers the specifics of each format and tips for preparing your data.

Supported formats

Format	Extensions	Parsing	Special features
CSV	.csv	Client-side (Papa Parse)	Auto-detects comma, tab, semicolon delimiters
Excel	.xlsx, .xls	Client-side	First sheet imported by default
SPSS	.sav	Server-side (pyreadstat)	Variable labels, value labels, metadata summary
Stata	.dta	Server-side	Variable labels as column headers (opt-in checkbox, on by default)

CSV files

CSV is the most universal format and the easiest to prepare. GraphHelix auto-detects the delimiter (comma, tab, or semicolon).

Requirements

• First row must be column headers
• One observation per row, one variable per column
• Consistent data types within each column (don't mix numbers and text in the same column)
• UTF-8 encoding recommended (most spreadsheet software exports this by default)

Tips

• Use short, descriptive column names without special characters (e.g., blood_pressure rather than Blood Pressure (mmHg) #1)
• Missing values can be empty cells or NA — GraphHelix handles both
• If your CSV has metadata rows above the header (common with instrument exports), remove them before importing

Excel files

GraphHelix imports the first sheet of .xlsx and .xls files. The same column structure rules apply as for CSV.

Common issues

• Merged cells — unmerge all cells before importing; merged cells cause misalignment
• Multiple header rows — only the first row is treated as column names
• Formulas — calculated values are imported correctly; formulas themselves are not
• Multiple sheets — only the first sheet is imported; move your data to Sheet 1 if needed

SPSS files (.sav)

SPSS files are parsed server-side and preserve the rich metadata that SPSS stores with your data.

What's preserved

• Variable labels — descriptive labels appear alongside variable names
• Value labels — coded values (1 = "Male", 2 = "Female") are shown in the import preview
• Variable count and row count — displayed in the import summary

Stata files (.dta)

Stata files are parsed server-side with full support for variable labels.

Variable labels

When importing a Stata file, you'll see a checkbox: "Use variable labels as column headers" (checked by default). This replaces short variable names (e.g., bp_sys) with their descriptive labels (e.g., "Systolic Blood Pressure"). A preview table shows the first 5 variable-to-label mappings so you can decide before importing.

Column type detection

GraphHelix automatically infers column types during import:

Detected type	Examples	Used for
Number	23.4, 100, -5.67	Statistical tests, continuous outcomes, predictors
Text	"Treatment", "Control", "Male"	Grouping variables, categorical factors
Date	2026-03-16, 03/16/2026	Time-to-event analysis, longitudinal data

You can override any column type during import if the auto-detection doesn't match your data.

Export options

After running analyses, you can export results in several formats:

• APA string — one-click copy of formatted result text
• Markdown — full analysis report with APA citation, summary, and AI interpretation
• JSON — complete results object for programmatic use
• Figures — PNG (150/300/600 DPI) or SVG with journal presets for Nature, Science, PLOS ONE, JAMA, and Cell
• Audit trail — Markdown table of all analyses for journal supplementary materials